The podcast addresses the challenge of context portability in the agentic AI era, proposing a solution: the personal context portfolio. This portfolio is a structured set of markdown files representing an individual's roles, projects, communication style, and expertise, serving as a machine-readable "operating manual" for AI agents. The portfolio consists of ten dimensions, each in its own markdown file, such as identity, roles and responsibilities, current projects, team relationships, and communication style. To facilitate creation, the host provides a public GitHub repo with templates and a personal context portfolio app that uses AI to interview the user and build the portfolio. The podcast further explores how to host this portfolio on an MCP server for easy access by various AI tools and agents.
Sign in to continue reading, translating and more.
Continue